【作者】 王鹏；

【导师】 庞辉；

【作者基本信息】 西安理工大学 ， 机械工程， 2022， 硕士

【摘要】 电动汽车作为缓解能源和环境问题、改善交通安全的理想解决方案,近几年来引起了众多学者们的研究兴趣,其中电动汽车的横摆稳定性是人们关注的热点问题。直接横摆力矩控制方法通过在电动汽车四个车轮上施加驱动力或制动力来生成附加横摆力矩,进而减小质心侧偏角和横摆率的输出响应,成为保证电动汽车横摆稳定性的重要控制策略。针对分布式驱动电动汽车的横摆稳定性控制问题,本文基于直接横摆力矩控制思想提出了一种电动汽车横摆稳定性分层协调控制策略,主要工作如下:(1)建立了电动汽车二自由度线性动力学模型和三自由度非线性动力学模型,为电动汽车行驶状态估计和横摆稳定性分层协调控制策略开发提供了可靠的模型基础。(2)在分布式驱动电动汽车三自由度动力学模型的基础上,搭建了基于无迹卡尔曼滤波算法(UKF)的汽车行驶状态估计方法。通过分析过程和测量噪声协方差对UKF算法估计车辆行驶状态精度的影响,总结了汽车行驶状态估计误差的变化规律。在此基础上建立了一种基于量纲一化新息平方的自适应噪声协方差调整策略,并将之和UKF算法相结合,提出了一种基于改进自适应无迹卡尔曼滤波算法(iAUKF)的车辆行驶状态估计方法,以提高汽车行驶状态估计的准确性和鲁棒性。通过CarSim和MATLAB-Simulink联合仿真对所提出的电动汽车状态估计方法进行了验证。(3)基于直接横摆力矩控制的思想,提出了一种电动汽车横摆稳定性分层协调控制策略。首先,利用所提出的基于UKF的汽车行驶状态估计方法实现对汽车质心侧偏角、横摆角速度以及纵、横向车速的精确估计。为了消除汽车参数不确定性对状态估计和横摆稳定性控制的影响,将汽车前后轮侧偏刚度作为可变参数进行了协同估计。在此基础上,以分布式驱动电动汽车二自由度线性动力学模型为基础搭建汽车横摆稳定性分层协调控制策略:上层附加横摆力矩控制器利用自适应反推算法设计了质心侧偏角和横摆角速度控制器,计算出汽车横摆稳定性控制所需的附加横摆力矩,下层最优转矩分配控制器以轮胎附着利用率为优化目标利用二次规划法实现附加横摆力矩在四个车轮轮毂电机上的附加扭矩分配。此外,为了消除自适应反推算法存在的复杂性爆炸问题,引入了一种改进的自适应神经动态表面控制(MANDSC)技术来对虚拟控制变量及其导数进行精确滤波,以减少分层控制策略的计算量、保证性能跟踪控制精度。在CarSim和MATLAB-Simulink环境下搭建了联合仿真测试平台,验证了所提电动汽车横摆稳定性分层协调控制策略的有效性。更多还原

【Abstract】 As an ideal solution to alleviate energy and environmental problems and improve traffic safety,electric vehicles(EVs)have attracted much research interest of many scholars in recent years,among which the yaw stability of EVs is the focus of people’s attention.Direct yaw moment control relies on the additional yaw moment generated by the drive or braking force applied to the four wheels to reduce the output response of sideslip angle and yaw rate,which has become an important control strategy to ensure the yaw stability of EVs.To solve the yaw stability control problem of the distributed drive EV,this paper proposed a hierarchical coordinated strategy for the yaw stability control based on the direct yaw moment control idea.The main works are summarized as follows:(1)The 2-DOF linear dynamics model and the 3-DOF nonlinear dynamics model of distributed drive EV are established,which provides a reliable control model for the accurate driving state estimation of the EVs,and the establishment of hierarchical coordinated control strategy of yaw stability for the EVs.(2)Based on the 3-DOF nonlinear dynamics model of the distributed drive EV,a vehicle driving state estimator based on unscented Kalman filter(UKF)is established.By analyzing the influence of process and measurement noise covariance on the estimation accuracy of UKF algorithm,the varying principles of estimation errors for vehicle driving states are summarized.Then,a normalized innovation square(NIS)-based adaptive noise covariance adaptive adjustment strategy is established and combined with the UKF algorithm to derive our improved adaptive unscented Kalman filter(iAUKF)-based vehicle driving state estimator.Finally,a comparative simulation investigation using CarSim and MATLAB-Simulink is conducted to validate the effectiveness and robustness of the proposed estimation method.(3)Based on the control idea of direct yaw moment,a hierarchical coordinated control strategy for yaw stability of EVs is proposed.First,a vehicle driving state estimation method based on UKF algorithm is proposed to accurately estimate the sideslip angle,yaw rate,longitudinal and lateral velocity of vehicle.In order to eliminate the influence of parameter uncertainty on state estimation and yaw stability control,the cornering stiffness coefficients of front and rear wheels are cooperatively estimated as the variable parameters.Next,a hierarchical coordinated control strategy consisting of the upper and lower controller is established based on the 2-DOF linear dynamics model of distributed drive EV to achieve the yaw stability.The upper layer controller is developed to generate the additional yaw moment,and taking the tire adhesion efficiency as the optimization objective,the lower layer optimal torque distribution controller is designed to realize the additional torque distribution of additional yaw moment on the hub motor of four wheels by using quadratic programming method.Moreover,in order to eliminate the complexity explosion existing in the adaptive backstepping method,a modified adaptive neural dynamic surface control(MANDSC)technique is introduced to accurately track the virtual control variables and their derivatives,so as to reduce the computation of hierarchical coordinated control strategy and ensure the performance tracking control.Finally,a co-simulation test platform is built in CarSim and MATLAB-Simulink environment to verify the effectiveness of the proposed hierarchical coordinated control strategy.更多还原